Many sociologists now assert that the long-term success of a society is dependent on its ability to socialize its young adult males -- in the sense of finding gainful employment for them in order to keep them busy and useful. Failure to do so -- for example, in inner cities in the United States, or in several African countries -- results in increased crime, civil unrest, etc. Apparently having a large number of testosterone-crazed individuals hanging around idle is a Bad Idea.
A follow-on thought for those who think that there will be lots of jobs programming the machines -- as far back as Heinlein's "The Moon is a Harsh Mistress" (copyright 1966), he recognized that given a sufficiently smart computer, it would be easier to tell it, "Mike, I need a program that does X, Y and Z" and let the machine write the program, as well as fill in the blanks for what else was implied by X, Y and Z, so the program did what it was supposed to.
I don't know if any group of laborers have, collectively, devoted more effort to putting themselves out of work than computer programmers.
Certainly this model has been used in a lot of science-fiction stories -- everyone gets a minimum stipend and lives on the productivity of the machines and a few humans. I've always wondered about the overall economic feasibility of such a system, and how we get from here to there. Here's a sample of questions -- many good stories have been written trying to answer some of them.
What standard of living will be produced? Will it be like I have now -- single-family house where everyone has their own bedroom, two cars, some amount of luxury? Or like things were when I was in college -- three people in a two-bedroom apartment, one car amongst them (and that 11 years old), no one ever ate out, etc? Or like rural India -- a family of five in a one-room hut?
Supply and demand have to balance in some fashion, so we need productivity to increase to the point that, say, 10% of the population plus the machines can produce enough goods and services to meet the demand. Some production should be simple, but will the humanoid robots be able to build or (more difficult) repair a house on site? Deal with the individual complexities of surgery? Teach small children? Settle a domestic dispute?
Who decides which people work and which don't? Are there tests? You're smart, you have to work and oversee the machines; your neighbor isn't, she gets to live the life of leisure. What if your neighbor wants to work, but has no skills that are useful?
Today, labor is taxed heavily and capital is taxed lightly -- look at how much of the total tax revenue derives from payroll and personal income taxes. You have to change to a system where capital is taxed heavily (your "taxes paid by those companies"), since there won't be enough labor to tax. Today's rich people are going to use some part of their money fighting that, and at least in the US, money appears to be quite persuasive.
Before anyone rushes out to try this, some basic questions about SCO's ownership arrangements:
In addition to the common stock, are there any special classes of shares? Some companies have "super-voting" classes where the special shares get 100 or 1000 votes per share. See, for example, Comcast prior to their deal with AT&T Broadband; the Roberts family held roughly 2% of the equity, but had over 80% of the votes.
What fraction of the shares are held by parties who may not be interested in selling at this time? If, for example, some small group already controls 51% of the shares and is willing to speculate on SCO winning the lawsuit (in which case those shares will be VERY valuable), they won't sell and no one else can gain control.
Are there any provisions in the corporate governance arrangements that require more than a simple majority? Perhaps changes to the board of directors require a super-majority?
There's no sense buying a share or ten unless you have a reasonable expectation that you can achieve the goal of controlling the company and making the changes that you want.
I, for one, do not want all of the devices in my home and/or office to be reachable by just anyone on the public Internet. I would be much happier with an architecture based on some sort of household gateway and a standard way for a device that needs external access to install an interpreted service proxy module on the gateway. The module can be as simple as port fowarding, or quite complex. But I want the added security that I believe an interpreted environment will provide. Modules written in Java would be fine.
IANATA (I am not a tax accountant), but I believe that, in the U.S., whether you get bit by this depends on what you do with the options. I think the tax bind goes like this:
You get 1,000 stock options with a strike price of $1.
At some point when the stock price is $51, you exercise the options and take the stock. You have an immediate taxable income of (51-1)*1000 = $50,000.
Later in the same year the stock and/or the market tanks and your shares go to say, $1 per share.
If your other income has pushed you into the top tax bracket (38%), you owe $19,000 in income tax on the option income. Even if you sell all of the stock, getting $1,000, you still have to come up with another $18,000 from somewhere to pay the taxes.
If your total income is high enough, you may also run into funky things with the Alternate Minimum Tax. Like many other past episodes involving U.S. taxes, the AMT is not indexed for inflation, so that a law intended to affect only the wealthy has, 20 years later, started biting the upper middle-class as well. At one point, some Senator said something like, "If we had voted for the kinds of tax rate increases that inflation and non-indexed tables have given us, none of us would have been re-elected."
I'm a chicken myself, and when I had the opportunity to exercise the small number of options that I had received, I used the "stockless" arrangement the company had set up -- the custodian sold my shares at the market price, subtracted the strike price from the proceeds, and sent me cash. It was all treated as straight income (no chance to exploit the lower capital gains rate), but no unexpected surprises at tax time either.
The National Science Foundation reports that China graduated nearly 200,000 engineers in 1999 from good universities that get better by the year. By comparison, American Universities graduate a mere 60,000 undergraduate engineers annually.
Combined, India and China produced nearly 26 percent of the world's newly minted engineers in 1999. Excluding Japan (where engineering wages are higher), Asian economies graduated 320,000 engineers in 1999 alone.
Like many other things -- steel, autos, DRAM -- the world now has the capacity to produce, and is producing, more engineers than it can efficiently use. Based on simple supply and demand, the price paid for engineers (wages) should fall.
When China built massive new steel production facilities, did they know that those mills could produce far more steel than China could use domestically? Without a doubt. But their cost structure is low and they planned to export the excess. Do they know that they don't need that many new engineers every year? Almost certainly, and they'll attempt to "solve" the problem the same way they solve their steel problem -- export.
Rather than seeking to control the supply of engineers, perhaps it would be interesting to ask how we might increase the demand for them, in both the developed and developing world.
but a lot of 'fancy' simple DTP tricks, such as embedded graphics with text that flows around them, it handles those poorly.
OTOH, both LaTeX and troff/groff (with appropriate macros) make floating displays almost effortless. A display is a collection of things, usually formatted text and graphics, that must be kept together. A floating display is one that should appear at a particular point in the document if there is room on the current page, but if not, it should "float" to the top of the next page while text continues to fill the current page. Plus, of course, rules for how to handle multiple displays that have been queued up and things like that.
A large complex academic paper or book goes through any number of iterations in which paragraphs and sections are added, removed or relocated. Having the software handle floating displays nicely is a big benefit. Floating displays have been in use in academia for a long time. Fifty years ago, my father worked his way through college doing typesetting at the local academic press. Once he overheard me berating Word for its lack of floating displays (circumstances had dictated that a particular paper had to be written in Word), and understood immediately what I was complaining about.
While not stated explicity, the article implies that the contracts include training for both users and admins, as well as consulting on the conversion from the current software (both OS and applications). Assuming a five-year contract, that $2600 per machine can also be thought of as $520 per year per machine for software, training and consulting. I expect the upfront consulting on moving the city's custom apps to the new OS is fairly expensive. And depending on the turnover rate of employees, per-year training may be a large program.
The I think the fact that the Suse/IBM initial quote was higher than the MS quote probably supports this view.
Very good. If I had moderator points today you'd get one.
Causality is always tricky, particularly in statistical "studies" like this. Maybe both effects (drop off in creative genius and marriage/kids) are the result of a single cause. Say, hypothetically, that a decrease in hormones below some magic threshold causes both (a) some change in brain chemistry that affects creativity and (b) exchange of the desire to screw everything in sight for the more practical results of staying home more and helping with the kids. Such a threshold could explain both the behavior in the article, as well as exceptions like Bach who remained creative and kept fathering kids for a long time -- maybe Bach was a mutant with a weird or nonexistant threshold.
It's not like we have a really good understanding of how brain chemistry really works...
Golf - While they have done a nice job reiging in club technology, you have ball technology going through the roof. Golf courses are being made obsolete. Expect the governing bodies to put in restrictions very soon to level things off.
IIRC, the fundamental limit set on balls by the USGA is that they shall not exceed an initial velocity of 250 feet per second (FPS) when hit on the USGA's standard hitting machine, and the machine hasn't changed in decades. There have been a variety of improvements in aerodynamics (ie, dimple design) that have improved things, but I don't think that's made an enormous difference. You've always been able to buy illegal "hot" balls, but the pros certainly aren't using them. Improvements in clubs -- compare the weight of a modern titanium-head graphite-shafted driver with an old maple-head steel-shafted one -- allow for higher clubhead velocities and resulting initial ball velocity. IIRC, modern pros get much higher initial velocities than the standard hitting machine, independent of ball technology.
For the average player, I believe that the biggest change has been the use of modern materials to lighten overall club weight and improve the distribution of head mass to produce a much larger "sweet" spot. That combination makes it easier to make solid contact with the ball, producing better energy transfer and greater average distance. OTOH, I don't think it has made the average player much better -- at least I haven't noticed that the average players' scores have improved dramatically.
Almost 30 years ago, I took a "free university" class in urban terrorism. It was basically a how-to taught by a disgruntled Vietnam vet. The two main lessons were (a) how to build large explosive devices on the cheap and (b) where to use them to best advantage. The 1995 Oklahoma City bombing used the same basic device he was teaching -- barrels full of high-nitrogen fertilizer and diesel fuel. As to where, his claim was that with between four and eight such devices, most medium-sized cities could be rendered uninhabitable because you could take out the electric power, water supply, sewer system, and telephone network. Locations of water works, phone company central offices, and such were all easily available at the public library.
Building the bombs is somewhat more difficult now than it was then. My brother-in-law works in a rural part of Kansas, and the fertilizer dealers are required to report unusual purchases to the FBI. I think some of the infrastructure is more robust now than it was then. Power companies have moved towards more grid-like architectures with multiple feeds serving an area. Widespread use of cell phones means there are more switches in more locations. OTOH, I live in a city of just about 100,000 with (I believe) only one water treatment facility and one sewage treatment facility.
Hmmm, most people who see the code say SCO's claims aren't FUD. Keep believing what you want, however, like everyone else on slashdot. Don't let facts, or lack of facts, get in the way.
Most of the people who see the code say there are anything from similarities to identical code (I find the fact that they don't all say there are identical blocks of code to be suspicious). TTBOMK, none of them say that SCO has shown any information about the history of said code, that is, evidence that would establish that the code was indeed SCO's and not lifted from some other source, and that the Linux code got there via IBM (the only company they're actually suing thus far). I expect proving the pedigree part will be difficult.
I think this is more true than most might want to admit. Once something comes into common usage, and people now expect it to work a particular way, it becomes more and more difficult to change. Consider the automobile -- there may be a terrific alternate UI that combines steering, acceleration and braking into a single intuitive form, but you'll have a hell of a time getting it adopted because too many people have too much invested in the current UI.
As well, the basic model of the Web is a collection of linked pieces of "smart" paper. There are only so many things you can do with all of those pages -- go forwards or backwards in the set of pages you've been perusing, or jump to something quite different by following a link. You can make the paper smarter (add more media types), or you can add some sort of search engine that's "always on" (how about a "next page like this one" button?), but how much more can you do with something that's still a collection of pages?
I'm not knocking the model -- books have survived for a long time because they're a really good way to organize and present information. I know that I can skim an index and the pages it references much quicker and easier than I could speak a phrase and listen to my computer list a set of pages or sections. And listening to it read the material while I decide if it's the right thing would drive me nuts.
And in many countries, a sizable tax income for the government. One of the reasons that international calls into India are so expensive is that the government charges a large tariff. IIRC, at one point taxes on incoming international calls made up almost 25% of the income of the Vietnamese government. Several years ago most countries signed a new agreement to get rid of many such tariffs. The developed countries had to eliminate them almost completely (with the immediate result of rates like 15 cents per minute from America to England) but undeveloped countries were allowed to keep their tariffs in place.
One thing that hasn't changed since the days of MSDOS is the underlying philosophy of Microsoft operating systems.
Bill Gates's vision of an operating system has always been that the os need be little more than a program launcher.
This assertion would seem worth discussing on at least two points.
First, there would seem to be a limited set of services that one would expect the OS (in the sense of the OS kernel) to provide: management of the hardware, loading and unloading programs, and some specialized functions like IP stack, inter-process communication, etc. But functions after the first two have more room for being optional -- many computers in many situations don't need network access, for example. Shouldn't any OS be primarily a good program loader/unloader?
Second, at least since Windows 95 and IE, Microsoft would seem to be at the forefront of putting new services and APIs into their operating system. Parts of the windowing system were moved into the OS kernel for performance purposes early on (as opposed to the UNIX/X approach, where the windowing system is always just a privileged application). Standard HTML-handling services and standard multimedia playback services are two that have gotten them in legal trouble, the first in the US antitrust case and the second in the ongoing EU antitrust investigations.
I believe you are correct, and that the Supreme Court ruled that all software was algorithms, and all algorithms were math, and math specifically cannot be patented. Since then, lower courts have held that some software, in binding together the parts of an apparatus or system, no single part being patentable by itself, was the innovative part of the overall system and could be patented.
The world has changed a lot in 20 years, and I think that there is a reasonable chance that the Supreme Court would uphold that reasoning in some cases; if not, you could have the situation of a physical device that performs a new function that is not patentable, but would be if the controlling "algorithm" were implemented in some other form. Why should an electronic circuit that implements a function using a new and complex feedback loop be patentable, but the same function implemented as software on a DSP not be? If the government can grant a temporary monopoly on a useful way to wire transistors and such together to do a job, why not a temporary monopoly on the use of a particular piece of code to do that same job?
That said, I'm afraid that many of the software patents that are granted by the PTO should not be because they fail the "novel" test. Donald Knuth said "I'm against patents on things that any student should be expected to discover." I would go further, and say that I'm against patents on things that competent programmers should be expected to discover. My name is on three software patents, but I believe that at least two of them should not have been granted because they fail the "novel" test. However, I worked for a giant corporation at the time that made a practice of obtaining software patents for defensive purposes.
IANAL, but -- if it infringes on the US patent, and you use it in the US, you're infringing. Patents give the holder complete control over the use of the invention for the period of the patent. Having an implementation done outside the US, which may be perfectly legal, doesn't gain you anything in terms of being able to use it legally inside the US.
I wonder if there is enough public funding for the US government to enter this new race. Compared to the last time they did it, (a) the total tax burden on the population (federal, state, and local) is already much higher, (b) the feds are already running very large deficits and (c) the feds are committed to massive future outlays in the social programs as the baby boom generation starts to retire.
I wonder if there's enough private funding for US private enterprise to enter this race. Assume the cost to develop the program is $100B (I'll bet it's more than that). What do the investors have when that's been spent? They want something that's returning on the order of $5B in profits to them per year. Has anyone seriously suggested any business opportunities that might produce profits on that scale?
Really, the meat of the SCO case revolves around the whole "derivative works" issue, and how sweeping the ownership/control claims they can make under such a contract.
This makes sense, and explains why it was "reasonable" for SCO to request an injunction shutting down all AIX users all over the world -- under their legal theory, IBM has no rights to AIX, only SCO does. A couple of random thoughts on arguing against SCO's position:
Before 1984, it was illegal for AT&T to sell an operating system. If the IBM contract predates that, it would seem reasonable to argue that neither AT&T nor anyone buying the contracts from them has any rights to monetary gain now.
In the USL and UC-Berkeley proceeding, I believe that judge had ruled that AT&T had improperly appropriated Berkeley code, not the other way 'round. Assuming the Berkeley and IBM licenses were similar, this might be a precedent against AT&T's rights to derivative works.
In some (not all) IP cases, failure to defend your rights in a timely fashion can cost you. The fact that no one has attempted to defend the rights to derivative works for the past several years may mean that those rights have "lapsed" in some sense.
For most people, anything approximating the functionality of "format c:" is going to seriously ruin their day/week/month. Or something that does fdisk-like manipulations of the partition table. Or starts at track 0 and writes random garbage over everything. For the large majority of owners, you don't have to actually damage the hardware in order to "break" the computer in ways that will require significant amounts of time and/or money to "repair".
Over the years, quite an assortment of viruses and worms have trashed hard disks.
Well, anyone whose native currency is the Euro, for example. If you decided to hold onto (hoard) dollars at the beginning of 2002, in the last 18 months your "money" has lost 30% of its value relative to holding Euros. Does this matter? At my local wine store, I've noticed that a couple of my favorite Italian vintages have increased more than 20% in price, while the California wines have stayed level.
I suppose it's possible that Microsoft and IBM and Oracle have something close enough to a monopoly in Europe that they've been able to increase the price (in local currencies) to offset the decrease in the value of the dollar. That's one of the indicators of a real monopoly -- the ability to change prices arbitrarily. Can anyone comment on MS product prices in Europe over the last 18 months? Does Office cost more now?
Yes, but...
the US is currently running a trade deficit with the UK -- we buy more stuff from you than you buy from us.
The dollars that the US spends on British products have to come back somehow, either directly or indirectly, unless you guys are just going to hoard them.
Given the recent slide of the dollar against most currencies, hoarding hasn't been a good idea.
Somewhere recently was an analysis that said the big difference was not in the percentage called strikes but the shape of the strike zone. The rules call for the width of the plate, knees to letters. What the author asserted was typically called was knees to beltline, from about an inch inside to about four inches outside, for some umpires up to six inches outside. The area of the two rectangles was approximately the same.
However, the author made the claim that a fastball belt-high two inches outside was easier to hit for power than a fastball letter-high over the middle of the plate, so the umpires' zone would result in an increase in extra-base hits and scoring. I don't know about that, but my own perception for the past several years is that, in the first couple of innings, control pitchers gradually pitch more and more outside while they establish how far outside the umpire's strike zone goes on that particular evening.
I guess I'm old fashioned; I like watching batters have to swing at high fastballs, and I like to watch pitchers bat.
I do. The standard has to meet several criteria in order to be "open". In particular, (1) it must be freely (as in free speech) available to anyone, (2) there must be no licensing or royalty requirements for anyone implementing it, and (3) such arrangements must be non-revocable. The standard will be considered preliminary until such time as (4) the specification has been available (see points 1-3) for a minumum of 90 days and (5) two independent implementations are shown to interoperate correctly. Discussions of changes to the standard by the issuing body (be it public or private) must be made public in a timely fashion.
Could MS (everyone here's favorite demon) issue an open standard for, say, an electronic document? Sure, so long as their discussions while creating the specification are public, the spec has been available with no restrictions to all developers for at least 90 days, and two independent implementations have been shown to interoperate. Could they change it arbitrarily in the future? Yes, but everyone would see it coming in time to be ready. Would they continue making such changes every six months? No, because it's more trouble than it's worth, even for them.
Slashdot Mirror
Another follow-on thought.
Many sociologists now assert that the long-term success of a society is dependent on its ability to socialize its young adult males -- in the sense of finding gainful employment for them in order to keep them busy and useful. Failure to do so -- for example, in inner cities in the United States, or in several African countries -- results in increased crime, civil unrest, etc. Apparently having a large number of testosterone-crazed individuals hanging around idle is a Bad Idea.
A follow-on thought for those who think that there will be lots of jobs programming the machines -- as far back as Heinlein's "The Moon is a Harsh Mistress" (copyright 1966), he recognized that given a sufficiently smart computer, it would be easier to tell it, "Mike, I need a program that does X, Y and Z" and let the machine write the program, as well as fill in the blanks for what else was implied by X, Y and Z, so the program did what it was supposed to.
I don't know if any group of laborers have, collectively, devoted more effort to putting themselves out of work than computer programmers.
Certainly this model has been used in a lot of science-fiction stories -- everyone gets a minimum stipend and lives on the productivity of the machines and a few humans. I've always wondered about the overall economic feasibility of such a system, and how we get from here to there. Here's a sample of questions -- many good stories have been written trying to answer some of them.
There's no sense buying a share or ten unless you have a reasonable expectation that you can achieve the goal of controlling the company and making the changes that you want.
I, for one, do not want all of the devices in my home and/or office to be reachable by just anyone on the public Internet. I would be much happier with an architecture based on some sort of household gateway and a standard way for a device that needs external access to install an interpreted service proxy module on the gateway. The module can be as simple as port fowarding, or quite complex. But I want the added security that I believe an interpreted environment will provide. Modules written in Java would be fine.
If your total income is high enough, you may also run into funky things with the Alternate Minimum Tax. Like many other past episodes involving U.S. taxes, the AMT is not indexed for inflation, so that a law intended to affect only the wealthy has, 20 years later, started biting the upper middle-class as well. At one point, some Senator said something like, "If we had voted for the kinds of tax rate increases that inflation and non-indexed tables have given us, none of us would have been re-elected."
I'm a chicken myself, and when I had the opportunity to exercise the small number of options that I had received, I used the "stockless" arrangement the company had set up -- the custodian sold my shares at the market price, subtracted the strike price from the proceeds, and sent me cash. It was all treated as straight income (no chance to exploit the lower capital gains rate), but no unexpected surprises at tax time either.
When China built massive new steel production facilities, did they know that those mills could produce far more steel than China could use domestically? Without a doubt. But their cost structure is low and they planned to export the excess. Do they know that they don't need that many new engineers every year? Almost certainly, and they'll attempt to "solve" the problem the same way they solve their steel problem -- export.
Rather than seeking to control the supply of engineers, perhaps it would be interesting to ask how we might increase the demand for them, in both the developed and developing world.
A large complex academic paper or book goes through any number of iterations in which paragraphs and sections are added, removed or relocated. Having the software handle floating displays nicely is a big benefit. Floating displays have been in use in academia for a long time. Fifty years ago, my father worked his way through college doing typesetting at the local academic press. Once he overheard me berating Word for its lack of floating displays (circumstances had dictated that a particular paper had to be written in Word), and understood immediately what I was complaining about.
The I think the fact that the Suse/IBM initial quote was higher than the MS quote probably supports this view.
Causality is always tricky, particularly in statistical "studies" like this. Maybe both effects (drop off in creative genius and marriage/kids) are the result of a single cause. Say, hypothetically, that a decrease in hormones below some magic threshold causes both (a) some change in brain chemistry that affects creativity and (b) exchange of the desire to screw everything in sight for the more practical results of staying home more and helping with the kids. Such a threshold could explain both the behavior in the article, as well as exceptions like Bach who remained creative and kept fathering kids for a long time -- maybe Bach was a mutant with a weird or nonexistant threshold.
It's not like we have a really good understanding of how brain chemistry really works...
For the average player, I believe that the biggest change has been the use of modern materials to lighten overall club weight and improve the distribution of head mass to produce a much larger "sweet" spot. That combination makes it easier to make solid contact with the ball, producing better energy transfer and greater average distance. OTOH, I don't think it has made the average player much better -- at least I haven't noticed that the average players' scores have improved dramatically.
Almost 30 years ago, I took a "free university" class in urban terrorism. It was basically a how-to taught by a disgruntled Vietnam vet. The two main lessons were (a) how to build large explosive devices on the cheap and (b) where to use them to best advantage. The 1995 Oklahoma City bombing used the same basic device he was teaching -- barrels full of high-nitrogen fertilizer and diesel fuel. As to where, his claim was that with between four and eight such devices, most medium-sized cities could be rendered uninhabitable because you could take out the electric power, water supply, sewer system, and telephone network. Locations of water works, phone company central offices, and such were all easily available at the public library.
Building the bombs is somewhat more difficult now than it was then. My brother-in-law works in a rural part of Kansas, and the fertilizer dealers are required to report unusual purchases to the FBI. I think some of the infrastructure is more robust now than it was then. Power companies have moved towards more grid-like architectures with multiple feeds serving an area. Widespread use of cell phones means there are more switches in more locations. OTOH, I live in a city of just about 100,000 with (I believe) only one water treatment facility and one sewage treatment facility.
I think this is more true than most might want to admit. Once something comes into common usage, and people now expect it to work a particular way, it becomes more and more difficult to change. Consider the automobile -- there may be a terrific alternate UI that combines steering, acceleration and braking into a single intuitive form, but you'll have a hell of a time getting it adopted because too many people have too much invested in the current UI.
As well, the basic model of the Web is a collection of linked pieces of "smart" paper. There are only so many things you can do with all of those pages -- go forwards or backwards in the set of pages you've been perusing, or jump to something quite different by following a link. You can make the paper smarter (add more media types), or you can add some sort of search engine that's "always on" (how about a "next page like this one" button?), but how much more can you do with something that's still a collection of pages?
I'm not knocking the model -- books have survived for a long time because they're a really good way to organize and present information. I know that I can skim an index and the pages it references much quicker and easier than I could speak a phrase and listen to my computer list a set of pages or sections. And listening to it read the material while I decide if it's the right thing would drive me nuts.
And in many countries, a sizable tax income for the government. One of the reasons that international calls into India are so expensive is that the government charges a large tariff. IIRC, at one point taxes on incoming international calls made up almost 25% of the income of the Vietnamese government. Several years ago most countries signed a new agreement to get rid of many such tariffs. The developed countries had to eliminate them almost completely (with the immediate result of rates like 15 cents per minute from America to England) but undeveloped countries were allowed to keep their tariffs in place.
First, there would seem to be a limited set of services that one would expect the OS (in the sense of the OS kernel) to provide: management of the hardware, loading and unloading programs, and some specialized functions like IP stack, inter-process communication, etc. But functions after the first two have more room for being optional -- many computers in many situations don't need network access, for example. Shouldn't any OS be primarily a good program loader/unloader?
Second, at least since Windows 95 and IE, Microsoft would seem to be at the forefront of putting new services and APIs into their operating system. Parts of the windowing system were moved into the OS kernel for performance purposes early on (as opposed to the UNIX/X approach, where the windowing system is always just a privileged application). Standard HTML-handling services and standard multimedia playback services are two that have gotten them in legal trouble, the first in the US antitrust case and the second in the ongoing EU antitrust investigations.
The world has changed a lot in 20 years, and I think that there is a reasonable chance that the Supreme Court would uphold that reasoning in some cases; if not, you could have the situation of a physical device that performs a new function that is not patentable, but would be if the controlling "algorithm" were implemented in some other form. Why should an electronic circuit that implements a function using a new and complex feedback loop be patentable, but the same function implemented as software on a DSP not be? If the government can grant a temporary monopoly on a useful way to wire transistors and such together to do a job, why not a temporary monopoly on the use of a particular piece of code to do that same job?
That said, I'm afraid that many of the software patents that are granted by the PTO should not be because they fail the "novel" test. Donald Knuth said "I'm against patents on things that any student should be expected to discover." I would go further, and say that I'm against patents on things that competent programmers should be expected to discover. My name is on three software patents, but I believe that at least two of them should not have been granted because they fail the "novel" test. However, I worked for a giant corporation at the time that made a practice of obtaining software patents for defensive purposes.
IANAL, but -- if it infringes on the US patent, and you use it in the US, you're infringing. Patents give the holder complete control over the use of the invention for the period of the patent. Having an implementation done outside the US, which may be perfectly legal, doesn't gain you anything in terms of being able to use it legally inside the US.
I wonder if there's enough private funding for US private enterprise to enter this race. Assume the cost to develop the program is $100B (I'll bet it's more than that). What do the investors have when that's been spent? They want something that's returning on the order of $5B in profits to them per year. Has anyone seriously suggested any business opportunities that might produce profits on that scale?
Sorry, I'm in a bad mood today...
Over the years, quite an assortment of viruses and worms have trashed hard disks.
Well, anyone whose native currency is the Euro, for example. If you decided to hold onto (hoard) dollars at the beginning of 2002, in the last 18 months your "money" has lost 30% of its value relative to holding Euros. Does this matter? At my local wine store, I've noticed that a couple of my favorite Italian vintages have increased more than 20% in price, while the California wines have stayed level.
I suppose it's possible that Microsoft and IBM and Oracle have something close enough to a monopoly in Europe that they've been able to increase the price (in local currencies) to offset the decrease in the value of the dollar. That's one of the indicators of a real monopoly -- the ability to change prices arbitrarily. Can anyone comment on MS product prices in Europe over the last 18 months? Does Office cost more now?
Yes, but... the US is currently running a trade deficit with the UK -- we buy more stuff from you than you buy from us. The dollars that the US spends on British products have to come back somehow, either directly or indirectly, unless you guys are just going to hoard them. Given the recent slide of the dollar against most currencies, hoarding hasn't been a good idea.
Somewhere recently was an analysis that said the big difference was not in the percentage called strikes but the shape of the strike zone. The rules call for the width of the plate, knees to letters. What the author asserted was typically called was knees to beltline, from about an inch inside to about four inches outside, for some umpires up to six inches outside. The area of the two rectangles was approximately the same.
However, the author made the claim that a fastball belt-high two inches outside was easier to hit for power than a fastball letter-high over the middle of the plate, so the umpires' zone would result in an increase in extra-base hits and scoring. I don't know about that, but my own perception for the past several years is that, in the first couple of innings, control pitchers gradually pitch more and more outside while they establish how far outside the umpire's strike zone goes on that particular evening.
I guess I'm old fashioned; I like watching batters have to swing at high fastballs, and I like to watch pitchers bat.
Could MS (everyone here's favorite demon) issue an open standard for, say, an electronic document? Sure, so long as their discussions while creating the specification are public, the spec has been available with no restrictions to all developers for at least 90 days, and two independent implementations have been shown to interoperate. Could they change it arbitrarily in the future? Yes, but everyone would see it coming in time to be ready. Would they continue making such changes every six months? No, because it's more trouble than it's worth, even for them.